The string2path R package converts a text to paths of the outlines of each glyph, based on a font data. Under the hood, this package is powered by the savvy framework to use these two Rust crates:
- ttf-parser for parsing font
data. TrueType font (
.ttf) and OpenType font (.otf) are supported. - lyon for tessellation of polygons and flattening the curves.
install.packages("string2path")The development version is available on R-universe:
install.packages("string2path", repos = c( yutannihilation = "https://yutannihilation.r-universe.dev", CRAN = "https://cloud.r-project.org" ) )
If you want to install from source, you need to have Rust toolchain installed before trying to install this package. See https://rust-lang.org/tools/install/ for the installation instructions.
library(string2path) library(ggplot2) d <- string2path("カラテが\n高まる。", "Noto Sans JP", font_weight = "bold") d <- tibble::rowid_to_column(d) ggplot(d) + geom_path( aes(x, y, group = path_id, colour = factor(glyph_id)), linewidth = 1.5 ) + theme_minimal() + coord_equal() + theme(legend.position = "top") + scale_colour_viridis_d(option = "H")
library(gganimate) d <- string2path("蹴", "Noto Sans JP") d <- tibble::rowid_to_column(d) ggplot(d) + geom_path( aes(x, y, group = path_id), linewidth = 2, colour = "purple2", lineend = "round" ) + theme_minimal() + coord_equal() + transition_reveal(rowid)
Note that "Noto Sans JP" above (and "Iosevka" below) is the font
installed on my local machine, so the same code might not run on your
environment. You can use dump_fontdb() to see the available
combination of font family (e.g. "Arial"), weight (e.g. "bold"), and
style (e.g. "italic").
dump_fontdb() #> # A tibble: 479 × 4 #> index family weight style #> <int> <chr> <dbl> <chr> #> 1 0 Arial 400 normal #> 2 0 Arial 400 italic #> 3 0 Arial 700 normal #> 4 0 Arial 700 italic #> 5 0 Arial 900 normal #> 6 0 Sitka Subheading 400 normal #> 7 0 Sitka Subheading 400 italic #> 8 0 等线 300 normal #> 9 0 等线 400 normal #> 10 0 等线 700 normal #> # ℹ 469 more rows
You can also specify the font file directly. Pomicons is a font available on gabrielelana/pomicons, licensed under SIL OFL 1.1.
pomicons_file <- here::here("fonts", "Pomicons.ttf") if (!file.exists(pomicons_file)) { dir.create(dirname(pomicons_file)) curl::curl_download( "https://github.com/gabrielelana/pomicons/blob/master/fonts/Pomicons.ttf?raw=true", destfile = pomicons_file ) } d_tmp <- string2path("\uE007", pomicons_file) ggplot(d_tmp) + geom_path(aes(x, y, group = path_id), linewidth = 5, colour = "#26d1a9") + theme_minimal() + coord_equal()
d <- string2fill( "abc", "Iosevka", font_weight = "bold", font_style = "italic" ) ggplot(d) + geom_polygon( aes(x, y, group = triangle_id, fill = factor(triangle_id %% 7)), colour = "grey", linewidth = 0.1 ) + theme_minimal() + coord_equal() + theme(legend.position = "none") + scale_fill_viridis_d(option = "H")
for (w in 1:9 * 0.01) { d <- string2stroke( "abc", "Iosevka", font_weight = "bold", font_style = "italic", line_width = w ) p <- ggplot(d) + geom_polygon( aes(x, y, group = triangle_id, fill = factor(triangle_id %% 2)), colour = "grey", linewidth = 0.1 ) + theme_minimal() + coord_equal() + theme(legend.position = "none") + scale_fill_manual(values = c("purple", "pink")) plot(p) }
tolerance controls resolution of the tessellation. You can reduce
tolerance to get higher resolutions. In most of the cases, 1e-5 ~
1e-6 should be enough. For more details, please refer to lyon’s
official
document.
for (tolerance in c(1e-1, 1e-2, 1e-3, 1e-4, 1e-5, 1e-6, 1e-7)) { d <- string2fill( "abc", "Iosevka", font_weight = "bold", font_style = "italic", tolerance = tolerance ) p <- ggplot(d) + geom_polygon( aes(x, y, group = triangle_id), fill = "transparent", colour = "black", linewidth = 0.5 ) + theme_minimal() + coord_equal() + ggtitle(paste0("tolerance: ", tolerance)) plot(p) }
